Home / News & Analysis / Pornhub Using AI to Identify, Tag Porn Stars

Pornhub Using AI to Identify, Tag Porn Stars

Finding videos from your favorite porn star on Pornhub is about to get a lot easier.

The porn site is now using artificial intelligence to detect and tag more than 10,000 porn stars in videos. The site previously relied on users and uploaders alone to manually tag them.

Pornhub fed the new AI model "several thousand videos," plus porn stars' official photos. Pornhub users can then validate the AI model's results by either upvoting or downvoting the tags. The company said the model will improve and get smarter as time goes on since it "learns based on what the community has validated."

"Now, users can search for a specific porn star they have an affinity for and we will be able to retrieve more precise results," Pornhub Vice President Corey Price said in a statement.

Related

The new model will "play a pivotal role moving forward," he added, since more than 10,000 videos are uploaded to the site every day. Pornhub has been testing the model in beta for the past month; during that time, it was able to scan 50,000 videos.

Over the next year, Pornhub hopes to scan all 5 million videos currently on its platform. The company also plans to start using this technology to tag specific categories and sexual positions featured in videos. So, if the AI model detects that a video is shot outside, it will automatically add it to the "Public" category for people who enjoy that type of thing. Or, if it detects a blonde in the video, it will add it to the "blonde" category.

"Artificial intelligence has quickly reached a fever pitch, with many companies incorporating its capabilities to considerably expedite antiquated processes," Price said. "And that's exactly what we're doing with the introduction of our AI model."

Read more

Check Also

Facebook mistakenly leaked developer analytics reports to testers

Set the “days without a Facebook’s privacy problem” counter to zero. This week, an alarmed developer contacted TechCrunch, informing us that their Facebook App Analytics weekly summary email had been delivered to someone outside their company. It contains sensitive business information including weekly average users, page views, and new users. 43 hours after we contacted Facebook about the issue, the social network now confirms to TechCrunch that 3 percent of apps using Facebook Analytics had their weekly summary reports sent to their app’s testers, instead of only the app’s developers, admins, and analysts. Testers are often people outside of a developer’s company. If the leaked info got to an app’s competitors, it could provide them an advantage. At least they weren’t allowed to click through to view more extensive historical analytics data on Facebook’s site. Facebook tells us it has fixed the problem and no personally identifiable information or contact info was improperly disclosed. It plans to notify all impacted developers about the leak today and has already begun. Below you can find the email the company is sending: Subject line: We recently resolved an error with your weekly summary email We wanted to let you know about a recent error where a summary e-mail from Facebook Analytics about your app was sent to testers of your app ‘[APP NAME WILL BE DYNAMICALLY INSERTED HERE]’. As you know, we send weekly summary emails to keep you up to date with some of your top-level metrics — these emails go to people you’ve identified as Admins, Analysts and Developers. You can also add Testers to your account, people designated by you to help test your apps when they’re in development. We mistakenly sent the last weekly email summary to your Testers, in addition to the usual group of Admins, Analysts and Developers who get updates. Testers were only able to see the high-level summary information in the email, and were not able to access any other account information; if they clicked “View Dashboard” they did not have access to any of your Facebook Analytics information. We apologize for the error and have made updates to prevent this from happening again. One affected developer told TechCrunch “Not sure why it would ever be appropriate to send business metrics to an app user. When I created my app (in beta) I added dozens of people as testers as it only meant they could login to the app…not access info!” They’re still waiting for the disclosure from Facebook. Facebook wouldn’t disclose a ballpark number of apps impacted by the error. Last year it announced 1 million apps, sites, and bots were on Facebook Analytics. However, this issue only affected apps, and only 3% of them. The mistake comes just weeks after a bug caused 14 million users’ Facebook status update composers to change their default privacy setting to public. And Facebook has had problems with misdelivering business information before. In 2014, Facebook accidentally sent advertisers receipts for other business’ ad campaigns, causing significant confusion. The company has also misreported metrics about Page reach and more on several occasions. Though user data didn’t leak and today’s issue isn’t as severe as others Facebook has dealt with, developers still consider their business metrics to be private, making this a breach of that privacy. While Facebook has been working diligently to patch app platform privacy holes since the Cambridge Analytica scandal, removing access to many APIs and strengthening human reviews of apps, issues like today’s make it hard to believe Facebook has a proper handle on the data of its 2 billion users.

Leave a Reply

Your email address will not be published. Required fields are marked *

Disclaimer: Trading in bitcoins or other digital currencies carries a high level of risk and can result in the total loss of the invested capital. theonlinetech.org does not provide investment advice, but only reflects its own opinion. Please ensure that if you trade or invest in bitcoins or other digital currencies (for example, investing in cloud mining services) you fully understand the risks involved! Please also note that some external links are affiliate links.